CN103968256B - Piping for tank farm leakage detection method - Google Patents
Piping for tank farm leakage detection method Download PDFInfo
- Publication number
- CN103968256B CN103968256B CN201410210893.2A CN201410210893A CN103968256B CN 103968256 B CN103968256 B CN 103968256B CN 201410210893 A CN201410210893 A CN 201410210893A CN 103968256 B CN103968256 B CN 103968256B
- Authority
- CN
- China
- Prior art keywords
- pressure
- pipeline
- leakage
- sequence
- transfinites
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Abstract
The present invention is a kind of combination assumed statistical inspection and the piping for tank farm leakage detection method of mode identification technology.The method utilizes oil depot monitoring system, Collecting operation field measuring instrument data and actuator state feedback information, determines that oil plant conveys working path, and pressure measuring instruments data are carried out with Kalman filtering, produces pressure newly to cease；Newly ceased using pressure, whether transfinited using many sequential probability ratio tests detection pipeline pressure, if assay transfinites, extract transfinite pressure sequence and its multidimensional temporal signatures；Identification model is leaked using the neutral net set up, pipe leakage and regulating working conditions are recognized.The method can detect pipeline pressure small change and determine whether pipeline leaks in the rate of false alarm and rate of failing to report of regulation with the shortest time.
Description
Technical field
The invention belongs to Discussion on Pipe Leakage Detection Technology field, it is related to statistical analysis technique and mode identification technology in oil depot pipe
Integrated application in road leak detection.
Background technology
Oil pipeline organically contacts defeated oil storage, loading and unloading oils installations and facilities as the essential pith of oil depot
Get up, oil depot oil transferring operating system is turned into an organic whole, so as to complete oil plant transmitting-receiving task.Piping for tank farm it is easy because
Connector seepage, pipe welding seam crackle and corrosion failure and leak.Because oil plant is inflammable, explosive, volatile and is easy to quiet
Electrofocusing, pipe leakage not only results in resource loss, reduces the fuel supply operational efficiency of oil depot, while environment is polluted, very
Fire explosion can extremely be caused, immeasurable loss is caused to oil depot.
At present, mainly have to the detection method of pipe leakage both at home and abroad leaking medium detection method, tube wall parameter detecting method and
Detection method based on software.
1. leaking medium detection method.The method with the air along pipeline, soil or water environment as detection object, according to it
In whether judge whether pipeline leaks along the change of thread environment containing oil or oil gas and pipeline, representative puts
Penetrating property tracer detection method, impingement, thermal infrared imaging method, cable leak detecting, fiber parametric amplification method etc..
2. tube wall parameter detecting method.The method records the burn into defect and weld seam shape of inner-walls of duct by detector in pipe
Condition, just can detect whether pipeline leaks by analyzer tube wall-like condition, including photodetection, video recording detection method, leakage field
Logical detection method, ultrasonic Detection Method, detection method of eddy and electromagnetic acoustic detection method etc..
3. the detection method of software is based on.The method is main by computer, with technologies such as signal transacting, pattern-recognitions
The physical parameters such as pipeline pressure, flow, temperature are analyzed with treatment to judge to leak and position leakage point, mainly including matter
Amount balancing method, pressure spot analytic approach, negative pressure wave method, pressure gradient method, acoustic-emission, Realtime Streaming Transport and statistical analysis technique
Deng.
Have that quantity is more, different in size, bore is small, flow is complicated, leakage rate is small, regulating working conditions are dry in view of piping for tank farm
The features such as disturbing many and sensor monitoring equipment and lack, thus piping for tank farm leak detection to the sensitivity of detection method, adaptability for working condition and
Versatility requirement is higher.Leaking medium detection method and tube wall parameter detecting method are required for additionally installing specific detection means, greatly
Continuous real-time detection cannot be realized more.Method based on software can be detected using existing pressure and flow sensor data
Leakage, the advantage with leak detection low cost, but for Small leak, slowly do not leak the relatively small parameter that causes and change insensitive, and not
Nominal situation regulation is adapted to, causes the generation of rate of false alarm higher.
The content of the invention
It is an object of the invention to provide a kind of combination assumed statistical inspection and mode identification technology, can be in regulation
In rate of false alarm and rate of failing to report, pipeline pressure small change is detected with the shortest time and the side whether pipeline leaks is determined
Method.
Piping for tank farm leakage detection method of the present invention is achieved through the following technical solutions：One kind is directed to oil depot
The leakage detection method of pipeline, the method utilizes oil depot monitoring system, pressure, temperature, flow, the liquid at collection POL storage operation scene
Measuring instrumentss data and the actuator state feedback informations such as position；According to gathered data and feedback information, the technique stream of oil depot is analyzed
Journey and work pattern, it is determined that current oil plant transport path, the pipeline pressure data of Collecting operation pipeline, and carry out Kalman filtering
Pressure is produced newly to cease；According to the rate of false alarm and rate of failing to report of regulation, newly ceased using pressure, using many sequential probability ratio test detection pipes
Whether road pressure transfinites, if assay transfinites, extracts transfinite pressure sequence and its multidimensional temporal signatures；Using what is set up
RBF neural leaks identification model, using the multidimensional temporal signatures of the pressure sequence that transfinites to be identified as mode input, enters one
Step Division identification pipe leakage and regulating working conditions；If RBF neural leaks identification model recognition result to leak, let out
Police is failed to report, pipeline pressure overrun testing and RBF neural leakage identification are continued if recognition result is regulating working conditions.
Piping for tank farm leakage detection method overall procedure of the present invention is as shown in Figure 1, it is characterised in that including following
Step：
Step 1：Using oil depot monitoring system, the measuring instrument such as pressure, temperature, flow, the liquid level at collection POL storage operation scene
Table data and actuator state feedback information；
Step 2：The technological process of the actual oil depot of analysis and work pattern, it is determined that current oil plant transport path, Collecting operation
The pipeline pressure data p of pipeline, pre-processes laggard line overrun and detects and extract the pressure sequence that transfinites, overrun testing flow such as Fig. 2
It is shown, including 5 specific steps：
Step 2-1：The pipeline pressure data of Collecting operation pipeline simultaneously carry out Kalman filtering, obtain pipeline pressure prediction
Value, calculates the difference of pipeline pressure and pipeline pressure predicted value, and after being standardized to it, generation standardizes pressure and newly ceases e；
Step 2-2：The distribution mean μ that standardization pressure newly ceases e is proposed to assume
H0:μ=0, H1:μ=θi
θiIt is alternative hvpothesis value, the prior distribution and average most short Check-Out Time principle according to different degrees of leakage determine many
Individual alternative hvpothesis value (θ1, θ2..., θm), m is alternative hvpothesis number, and alternatively θ is assumed to be after calculating k samplingiIt is sequential general
Rate is than inspection (SPRT：Sequential Probability Ratio Test) statistic Yi(k)：
The standardization pressure generated after e (j) samples for j times in formula newly ceases, 1≤j≤k；
Step 2-3：Compensation rate U amendment detection statistics are introduced, remembers that revised test statistics is Yi' (k), calculate public
Formula is as follows：
Yi' (k)=Yi(k)+U i=1,2 ..., m
Step 2-4：Transfinite judgement, if the rate of false alarm of inspection provision is set to α, rate of failing to report is set to β, then transfinite judgement threshold of detectability
Value B so determines：B=ln [(1- β)/α]；The test statistics that any alternative is assumed exceedes inspection threshold value, i.e. max { Y1'(k),
Y′2(k),…,Y′m(k) } >=B when, judge that pipeline pressure transfinites, go to step 2-5；As max { Y1'(k),Y′2(k),…,Y′m
(k)}<During B, going to step 2-1 carries out (k+1) secondary sampling；
Step 2-5：Extraction is transfinited pressure sequence；Assuming that during k moment, Y 'n(k)=max { Y1'(k),Y′2(k),…,Y′m
(k) } >=B, 1≤n≤m, then to Y 'nK () is analyzed, if Y 'nK () meets
Namely since the r moment Y 'nK () is continuously more than zero, then estimate to start exception in r moment pipeline pressures, from pressure
Abnormal starting point to the sequence length for detecting pressure limit is (k-r), and it is 2 to extract (2r-k) moment between the k moment, length
(k-r) reset pressure sequence piIt is the pressure sequence that transfinites, 2r-k≤i≤k；
Step 3：The pressure sequence that transfinites further is analyzed, pipe leakage and regulating working conditions, including 3 specific steps are distinguished：
Step 3-1：The pressure sequence that will transfinite is normalized, and calculates the average amplitude t of the sequence1, root width
Value t2, variance t3, root mean square t4, kurtosis t5：
Form multidimensional time domain charactreristic parameter collection T={ t1,t2,t3,t4,t5}
Step 3-2：Set up RBF (Radial Basis Function, RBF) neutral net leakage identification mould
Type, mode input nodes are temporal signatures dimension, and output node number is conduit running state encoding dimension, with training sample
Temporal signatures collection T={ t1,t2,t3,t4,t5}lAs input, correspondence pipeline running status as desired output, complete leakage and know
The training of other model, wherein l is training sample group number；
Step 3-3：By the multidimensional time domain charactreristic parameter collection T={ t of the sequence that transfinites to be identified1,t2,t3,t4,t5Input
RBF neural leaks identification model, judges pipeline in regulating working conditions state according to output or leaks；
Step 4：If RBF neural leaks identification model recognition result to leak, leakage alarm；If identification knot
Fruit is then transferred to step 1 for regulating working conditions.
The beneficial effects of the invention are as follows：The pipeline pressure data of oil depot monitoring system collection are taken full advantage of, except existing
Outside pipe pressure sensor, without buying and installing special leakage detection apparatus in addition；Sequential probability ratio test method is improved, is made
Pipeline pressure small change can be detected with the shortest time with many sequential probability ratio test methods；In pipeline pressure overrun testing base
On plinth, pipe leakage and regulating working conditions are recognized using the mode identification method of RBF neural, reduce the wrong report of leak detection
Rate, improves the working conditions change adaptability of leak hunting method.
Brief description of the drawings
Fig. 1 is piping for tank farm leak detection flow；
Fig. 2 is pressure limit testing process, and SPRT1, SPRT i, SPRT m represent alternative hvpothesis θ respectively in figure1、θi、θm
Sequential probability ratio test；Y1, Yi, Ym are respectively alternative hvpothesis θ1, θi, θmSequential probability ratio test statistic；Y1'、Yi'、
Y′mRespectively alternative hvpothesis θ1, θi, θmRevised sequential probability ratio test statistic；
Fig. 3 is pipeline pressure sequence and standardization pressure innovation sequence；
Fig. 4 is pipeline pressure overrun testing curve；
Fig. 5 transfinites pressure sequence diagram for extraction.
Specific embodiment
Embodiment lifts this example to illustrate specific embodiment of the invention.Analysis oil house processes flow and currently work first
Industry pattern, the pressure data p of Collecting operation pipeline, generation pressure newly ceases e, pipeline pressure and standardization after carrying out Kalman filtering
Pressure innovation sequence is as shown in Figure 3.
The distribution mean μ that standardization pressure newly ceases e is proposed to assume
H0:μ=0, H1:μ=θi
θiIt is alternative hvpothesis value, 0.1,0.2,0.3,0.4 is respectively in embodiment, calculates alternative hvpothesis θ after sampling every timei
Test statistics YiK () is as follows：
Calculate the modified survey statistic Y after introducing compensation rate Ui'(k)：
Yi' (k)=Yi(k)+U i=1,2 ..., m
Specify inspection rate of false alarm α=0.05, rate of failing to report β=0.05, then check threshold value B=ln [(1- β)/α]=
2.944, as max { Y1'(k),Y′2(k),…,Y′m(k) } >=B when, judge that pipeline pressure transfinites；As max { Y1'(k),Y′2
(k),…,Y′m(k)}<Continue to sample and calculate test statistics during B and compare with threshold value B.Fig. 4 is alternative hvpothesis point in embodiment
Not Wei 0.1,0.2,0.3,0.4 sequential probability ratio test curve, after the 368th sampling it is alternative be assumed to be 0.3 inspection statistics
Amount Y3' exceed threshold value B at first, detect that pipeline pressure transfinites.
It is that 0.3 test statistics is analyzed to alternative hvpothesis, as shown in figure 5, the 368th sampling instant is super for pressure
Point of accumulation k=368；Since 309 sampling instants, Y3' start continuously to be more than zero, r=309, it is continuously 59 more than zero-length；Carry
It is the pressure sequence that transfinites of 2 (k-r) to take (2r-k) moment between the k moment, length, that is, extract from the 250th sampling instant to 368
The reset pressure sequence of secondary sampling instant is used as the pressure sequence that transfinites.
The multidimensional temporal signatures of training sample are calculated, the training sample sequence temporal signatures result that transfinites is as shown in table 1.
The training sample of table 1 transfinites sequence temporal signatures
As shown in table 2, running status correspondence 3 is encoded conduit running state encoding.
The conduit running state encoding of table 2
| Conduit running state | State encoding |
| Adjust valve | 100 |
| Leakage | 010 |
| Adjust pump | 001 |
RBF neural leakage identification model is set up according to table 1, the data of table 2.
By the pressure sequence temporal signatures input RBF neural leakage identification model that transfinites to be detected, known according to output
No matter road running status, pipe leakage recognition result is correct, as shown in table 3.
The pipe leakage recognition result of table 3
Claims (1)
1. a kind of leakage detection method for piping for tank farm, it is characterised in that the detection method step is as follows：
Step 1, using oil depot monitoring system, the measuring instrumentss number such as pressure, temperature, flow, the liquid level at collection POL storage operation scene
According to and actuator state feedback information；
Step 2, the technological process for analyzing actual oil depot and work pattern, it is determined that current oil plant transport path, Collecting operation pipeline
Pipeline pressure data p, pre-process laggard line overrun and detect and extract the pressure sequence that transfinites, that is to say the pipe of Collecting operation pipeline
Road pressure data simultaneously carries out Kalman filtering, obtains pipeline pressure predicted value, calculate pipeline pressure and pipeline pressure predicted value it
Difference, and after being standardized to it, generation standardization pressure newly ceases e；The distribution mean μ that standardization pressure newly ceases e is proposed false
If H0:μ=0, H1:μ=θi, θiIt is alternative hvpothesis value, prior distribution and average most short Check-Out Time according to different degrees of leakage
Principle determines multiple alternative hvpothesis value (θ1, θ2..., θm), m is alternative hvpothesis number, alternative hvpothesis after k sampling of calculating
Sequential probability ratio test statistic Y '1(k),Y′2(k),…,Y′m(k)；The test statistics that any alternative is assumed exceedes threshold of detectability
Value, i.e. max { Y '1(k),Y′2(k),…,Y′m(k) } >=B when, judge that pipeline pressure transfinites；Assuming that during k moment,1≤n≤m, then to Y 'nK () is analyzed, if Y 'nK () meets
Namely since the r moment Y 'nK () is continuously more than zero, then estimate to start exception in r moment pipeline pressures, is opened from pressure anomaly
Initial point to the sequence length for detecting pressure limit is (k-r), and it is 2 (k-r) to extract (2r-k) moment between the k moment, length
Reset pressure sequence piIt is the pressure sequence that transfinites, 2r-k≤i≤k；
Step 3, further analysis are transfinited pressure sequence, distinguish pipe leakage and regulating working conditions；
If step 4, RBF neural leak identification model recognition result to leak, leakage alarm；If recognition result is
Regulating working conditions are then transferred to step 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410210893.2A CN103968256B (en) | 2014-05-19 | 2014-05-19 | Piping for tank farm leakage detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410210893.2A CN103968256B (en) | 2014-05-19 | 2014-05-19 | Piping for tank farm leakage detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103968256A CN103968256A (en) | 2014-08-06 |
| CN103968256B true CN103968256B (en) | 2017-06-06 |
Family
ID=51238092
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410210893.2A Expired - Fee Related CN103968256B (en) | 2014-05-19 | 2014-05-19 | Piping for tank farm leakage detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103968256B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105156905A (en) * | 2015-07-09 | 2015-12-16 | 南京声宏毅霆网络科技有限公司 | Leakage monitoring system, method and device for pipeline and server |
| US20180292292A1 (en) * | 2015-11-04 | 2018-10-11 | Nec Corporation | Pipe condition detection device, pipe condition detection method, computer-readable recording medium, and pipe condition detection system |
| CN106090630B (en) * | 2016-06-16 | 2018-07-31 | 厦门数析信息科技有限公司 | Fluid pipeline leak hunting method based on integrated classifier and its system |
| CN106125661A (en) * | 2016-07-15 | 2016-11-16 | 上海海事大学 | Assembly shop fuel transfer system pipeline PLC monitoring system |
| CN106594526B (en) * | 2016-12-02 | 2019-04-12 | 上海环宇消防集团有限公司 | A kind of water supply network state monitoring method and device based on hydraulic pressure sampled data |
| CN106870958B (en) * | 2017-03-22 | 2019-04-02 | 上海伟梦物联网科技有限公司 | A kind of detection of water pressure condition and transfer method and its realization device |
| CN108050397B (en) * | 2018-01-24 | 2019-04-05 | 东北大学 | Layering based on optical fiber source signal is sequential than pipe leakage monitoring method and system |
| CN109058771B (en) * | 2018-10-09 | 2019-11-12 | 东北大学 | The pipeline method for detecting abnormality of Markov feature is generated and is spaced based on sample |
| CN109340582A (en) * | 2018-11-22 | 2019-02-15 | 深圳市欧佩亚海洋工程有限公司 | A kind of submarine pipeline leakage monitoring method and system |
| CN112842277B (en) * | 2021-02-08 | 2022-08-09 | 上海理工大学 | Fall detection method and device based on multiple sequential probability ratio detection |
| CN113076817B (en) * | 2021-03-17 | 2022-11-04 | 上海展湾信息科技有限公司 | Weld pore defect real-time detection method and system |
| CN116382100B (en) * | 2023-06-05 | 2023-08-01 | 中特检管道工程（北京）有限公司 | Oil and gas pipeline detection control system and control method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1184931A (en) * | 1996-12-11 | 1998-06-17 | 唐秀家 | Method and apparatus for detecting and positioning leakage of fluid transferring pipeline |
| US6513542B1 (en) * | 2000-08-08 | 2003-02-04 | Taiwan Semiconductor Manufacturing Co., Ltd | Liquid supply or drain pipe equipped with a leakage detector |
| CN1584531A (en) * | 2004-05-31 | 2005-02-23 | 天津大学 | Method for detecting leakage of oil-gas pipeline based on pressure signals |
| CN1755342A (en) * | 2004-09-28 | 2006-04-05 | 北京埃德尔黛威新技术有限公司 | Method and apparatus for detecting leakage of liquid pressure pipeline |
-
2014
- 2014-05-19 CN CN201410210893.2A patent/CN103968256B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1184931A (en) * | 1996-12-11 | 1998-06-17 | 唐秀家 | Method and apparatus for detecting and positioning leakage of fluid transferring pipeline |
| US6513542B1 (en) * | 2000-08-08 | 2003-02-04 | Taiwan Semiconductor Manufacturing Co., Ltd | Liquid supply or drain pipe equipped with a leakage detector |
| CN1584531A (en) * | 2004-05-31 | 2005-02-23 | 天津大学 | Method for detecting leakage of oil-gas pipeline based on pressure signals |
| CN1755342A (en) * | 2004-09-28 | 2006-04-05 | 北京埃德尔黛威新技术有限公司 | Method and apparatus for detecting leakage of liquid pressure pipeline |
Non-Patent Citations (2)
| Title |
|---|
| 卡尔曼滤波和序贯概率比检验在管道泄漏监测中的应用;朱爱华;《中国优秀博硕士学位论文全文数据库(硕士)工程科技II辑》;20070515(第5期);第8-10页、第16-18页、第25-26页、第34页 * |
| 原油管道泄漏检测与定位技术的研究;梁斌;《中国优秀硕士学位论文全文数据库工程科技I辑》;20100315(第3期);第18页、第25-28页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103968256A (en) | 2014-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103968256B (en) | Piping for tank farm leakage detection method | |
| Lu et al. | Leakage detection techniques for oil and gas pipelines: State-of-the-art | |
| Murvay et al. | A survey on gas leak detection and localization techniques | |
| Boaz et al. | An overview of pipeline leak detection and location systems | |
| Song et al. | Gas leak detection in galvanised steel pipe with internal flow noise using convolutional neural network | |
| Liu et al. | Fast leak detection and location of gas pipelines based on an adaptive particle filter | |
| CA2960587C (en) | Device and method for fluid leakage detection in pressurized pipes | |
| Lang et al. | A multiple leaks’ localization method in a pipeline based on change in the sound velocity | |
| CA2409886C (en) | Duct leakage control | |
| NO335721B1 (en) | Method for direct detection of leakage in emergency shut-off valves or other valves | |
| CN103807607A (en) | Novel device and method for detecting leakage of fluids in pipelines | |
| Fiedler | An overview of pipeline leak detection technologies | |
| CN105890844A (en) | Qualitative and quantitative detection method for tiny leakage of hidden oil storage tank | |
| Xue et al. | Application of acoustic intelligent leak detection in an urban water supply pipe network | |
| Wang et al. | Leak detection in pipeline systems using hydraulic methods: A review | |
| Idachaba et al. | Current technologies and the applications of data analytics for crude oil leak detection in surface pipelines | |
| Anastasopoulos et al. | ACOUSTIC EMISSION LEAK DETECTION OF LIQUID FILLED BURIED PIPELINE. | |
| Mandal | Gas leak detection in pipelines & repairing system of titas gas | |
| CN103032626A (en) | System and method for diagnosing fault of adjusting valve | |
| Shama et al. | Review of leakage detection methods for subsea pipeline | |
| CN207455197U (en) | Pipeline leakage testing positioning experiment system | |
| Chuka et al. | Transient model-based leak detection and localization technique for crude oil pipelines: a case of NPDC, Olomoro | |
| Shehadeh et al. | Modelling the effect of incompressible leakage patterns on rupture area in pipeline | |
| Xu et al. | Leak detection methods overview and summary | |
| US20220196512A1 (en) | Detection of a leakage in a supply grid |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170606 Termination date: 20180519 |